An approximate dynamic programming approach to admission control in WCDMA networks

This paper presents a Connection Admission Control (CAC) algorithm for Wide-CDMA (WCDMA) networks based on an Approximate Dynamic Programming (ADP) approach: the network is modeled as a Markov Decision Process (MDP), and ADP algorithms are used to compute the optimal admission policy. Two innovative concept are introduced: i) by formulating the problem as a linear programming problem, the blocking probabilities of the different classes of service supported by the network are controlled by appropriately modifying the reward function; ii) on the ground of practical considerations on the CAC problem, a restricted structure policy is determined, based on the reduction of the action space for low load conditions. Numerical simulation results show the effectiveness of the approach.